When AI Customer Service Goes Wrong: My Experience with Commonwealth Bank

AI customer service is supposed to make life easier. But what happens when automation makes it harder for someone to become a customer?

I recently had an experience with Commonwealth Bank of Australia (CommBank) that perfectly illustrates how automated customer service can go wrong, frustrate potential customers and create unnecessary barriers to doing business.

All I wanted to do was become a new customer of Commonwealth Bank.

A fairly straightforward request, you would think.

Instead, I found myself going around in circles with an automated telephone system that repeatedly asked me to provide a PIN or banking password for an account I didn’t have.

It was frustrating, confusing and, from a marketing and customer experience perspective, a remarkable example of how technology can get in the way of winning new business.

There are some valuable lessons here for any business considering AI customer service, automated phone systems or AI-powered customer engagement.

My CommBank Customer Service Experience: I Just Wanted to Become a Customer

On 21 November 2026, I called Commonwealth Bank to enquire about becoming a new customer.

I wasn’t calling to complain about an existing account. I wasn’t trying to recover a password or resolve a banking problem.

I wanted to give them my business.

After listening to an introductory announcement about changes to home loan interest rates, the automated customer service system asked me a simple question:

“So tell me, how can we help you today?”

My response was equally simple.

“I want to become a new customer.”

The system acknowledged my request before asking whether my enquiry related to an existing account.

No. I wanted to become a new customer.

Next, it asked whether I wanted to open a Youth Saver account.

Again, no.

Then things became particularly frustrating.

The system asked me to enter my CommBank app PIN or phone banking password to verify my identity.

There was just one problem.

I wasn’t a CommBank customer. I didn’t have either of those things.

I explained this repeatedly, but the automated system continued asking for the same information.

At one point, I stated:

“I don’t have that. I want to become a new customer of Commonwealth Bank of Australia.”

Yet the system returned to its standard verification instructions.

It was like being stuck in a conversation with someone who could hear the words I was saying but couldn’t understand what they meant.

After several attempts, the interaction still hadn’t provided a straightforward path to becoming a customer.

And that raises a much bigger question.

How Many Potential Customers Are Businesses Losing Through Poor Automation?

From a marketing perspective, what happened is particularly interesting.

Businesses invest considerable amounts of money attracting potential customers.

They spend on Google Ads, SEO, social media marketing, brand awareness, website development and lead generation.

The objective is to convince someone that they should consider doing business with them.

But imagine spending all that money to generate interest, only to create barriers when someone actually wants to become a customer.

That’s the problem with a poor customer experience.

Customer acquisition doesn’t finish when someone makes an enquiry. It finishes when you’ve successfully converted that interest into a customer.

In my case, I was already interested in becoming a Commonwealth Bank customer. The bank didn’t need to persuade me to make contact.

All it needed to do was help me take the next step.

Instead, the automated phone system added friction to what should have been an easy process.

Now, I’m not suggesting one frustrating telephone interaction proves CommBank is losing a particular number of customers.

But it does illustrate a genuine commercial risk.

If someone has to work too hard to give you their business, they may simply go somewhere else.

And in a competitive market, making it easy to become a customer should be fundamental.

Is AI Customer Service Actually Improving the Customer Experience?

Artificial intelligence has enormous potential to improve customer service.

AI chatbots, virtual assistants and automated telephone systems can help businesses respond to enquiries, answer questions and connect customers with the information they need.

When implemented well, automation can reduce waiting times, improve service availability and make interactions more convenient.

But there’s an important distinction between automating a process and improving a customer experience.

Automation isn’t automatically an improvement.

In my CommBank experience, the system appeared to follow a verification process designed for existing customers, even though I had repeatedly explained I wasn’t one.

I haven’t independently verified whether the particular system handling my call used generative AI, traditional voice recognition or another type of automated call routing.

But regardless of the technology, the underlying customer experience problem is the same.

The system wasn’t responding effectively to my stated intention.

It was asking me to complete a step I couldn’t complete.

And rather than finding another path forward, it kept repeating the same instructions.

That’s not the kind of experience businesses should be trying to create.

Five Lessons for Businesses Using AI in Customer Service

Whether you’re a large corporation or a small Australian business, my experience highlights five important considerations when introducing AI or customer service automation.

1. Understand What Your Customer Actually Wants

The first priority of an effective customer service system should be understanding the customer’s intent.

Are they an existing customer looking for assistance? Are they making a complaint? Do they want information about a product? Or are they a prospective customer ready to buy?

These are very different journeys.

A potential new customer shouldn’t automatically be pushed through the same verification process as an existing account holder.

In my case, the words “I want to become a new customer” should have been enough to initiate a relevant conversation or route me to the appropriate team.

Customer intent should determine the journey, not the other way around.

2. Make It Easy for People to Give You Their Business

One of the most important principles of marketing is removing barriers to conversion.

Think about what makes a high-converting website effective.

Clear messaging. Simple navigation. Obvious calls to action. Easy enquiry forms. Helpful information.

These same principles should apply to telephone customer service and AI-powered conversations.

When someone wants to become a customer, don’t make the process unnecessarily difficult.

Welcome them. Understand their needs. Guide them towards the next step.

The easier you make the experience, the better your chances of converting that opportunity.

3. Always Provide a Way Out of an Automation Loop

One of the most frustrating aspects of my experience was repeatedly being asked for information I couldn’t provide.

A well-designed automated system should recognise when an interaction isn’t progressing.

If someone has already explained that they don’t have a required password, repeating the same request is unlikely to achieve a different outcome.

There should be an alternative.

That might mean offering another verification method, transferring the enquiry to an appropriate team or allowing the person to speak with a human representative.

AI should support customer service, not become a barrier between customers and the people who can help them.

4. Humanise the AI Customer Experience

Businesses can become so focused on what technology can do that they forget how customers want to feel.

People want to be understood.

They want to feel welcomed, especially when they’re considering becoming a new customer.

They want straightforward answers and confidence that they’re dealing with a business that values their time.

A friendly voice or conversational chatbot interface doesn’t automatically create a humanised experience.

The real measure is whether the system responds appropriately to what someone is trying to achieve.

If someone says they want to become a customer, a welcoming and useful response should follow.

Technology should make the relationship feel easier, not more mechanical.

5. Test Your AI Customer Service Before and After Launch

Perhaps the biggest lesson from this experience is the importance of testing.

When I finished the call, I couldn’t help wondering how this particular customer journey had been tested.

Had someone attempted to navigate the system as a prospective customer?

Had they tried calling without an existing account, PIN or password?

Had they tested what happens when someone repeatedly tells the system that its instructions don’t apply?

I don’t know what testing CommBank conducted, but the experience raises important questions that every business should be asking.

AI and automated customer service systems need to be tested against real customer scenarios.

Not just the ideal journey where everything goes according to plan.

They need to be tested against misunderstandings, unexpected answers, frustrated customers and people who don’t fit the default workflow.

And that testing shouldn’t stop once the system goes live.

Businesses should continually review customer interactions, identify where people are getting stuck and improve the experience.

The Hidden Marketing Cost of Poor Customer Service

It’s easy to think of customer service and marketing as separate functions.

They’re not.

Every interaction someone has with your business contributes to how they perceive your brand.

Your advertising might promise convenience, simplicity and outstanding service.

But if the actual experience is confusing or difficult, that promise quickly loses credibility.

Consider what happened during my interaction with CommBank.

I began the call with a positive intention: becoming a new customer.

By the end, I was frustrated and questioning why the process was so difficult.

That change in perception happened during a single customer service interaction.

For businesses, this illustrates why the entire customer journey matters.

Your marketing can attract attention and generate enquiries, but your customer experience ultimately helps determine whether people choose to do business with you.

A great marketing campaign can’t consistently compensate for a poor customer experience.

AI Isn’t the Problem. Poor Customer Experience Design Is.

I don’t believe businesses should avoid AI because some automated systems deliver frustrating experiences.

Quite the opposite.

AI presents enormous opportunities to improve marketing, customer service and business operations.

But businesses need to approach implementation with a clear understanding of what they’re trying to achieve.

Are you genuinely improving the customer experience?

Are you making it easier for customers to access your products and services?

Are you helping people find the information they need more quickly?

Or are you simply introducing automation because the technology is available?

There’s a big difference.

A successful AI implementation should be measured by the outcomes it delivers for customers and the business, not simply how many interactions it automates.

My Final Thoughts on Commonwealth Bank’s Automated Customer Service

My experience calling Commonwealth Bank was disappointing.

I wanted to become a new customer, yet the automated system repeatedly requested credentials that I couldn’t possibly have as someone who hadn’t joined the bank.

I hope Commonwealth Bank uses experiences like this to identify opportunities to improve its customer journey.

But the lessons extend well beyond one bank.

Every business introducing AI chatbots, automated customer service or digital self-service tools needs to think carefully about the people using them.

Start with the customer. Understand what they need. Design a simple journey. Test it thoroughly. And always make it possible to get help when the technology isn’t working.

Because ultimately, the purpose of marketing and customer service hasn’t changed.

Make it easy for people to understand what you offer, trust your business and become a customer.

AI should help achieve that.

Not get in the way.

Have You Experienced a Frustrating AI Customer Service Bot?

I’d be interested to hear whether you’ve had a similar experience with AI chatbots, automated telephone systems or digital customer service.

Have these technologies made dealing with businesses easier, or have they added unnecessary frustration?

Share your experience in the comments.

And if you’re a business owner implementing AI, here’s a simple challenge:

Try becoming a new customer of your own business.

You might be surprised by what you discover.

Everything I write about is my own opinion or things I’ve either researched, taken a picture of, seen news about, and want to share. Let’s keep the conversation going, post a comment below.

Leave a Reply

Your email address will not be published. Required fields are marked *

All fields marked with * must be filled.
Please enter a valid email.